JP2019168514A - Map data creation device - Google Patents

Abstract

To create three-dimensional map data that is high in accuracy.SOLUTION: A map data creation device is configured to create map data for a three-dimensional map of an object area. The map data creation device comprises: an environment information storage unit that stores K pieces (K is a natural number equal to or more than one) of environment information on each of K pieces of measurement points in the object area; an inclination information storage unit that stores a first inclination with respect to a horizontal surface from an N-1-th measurement point to an N-th measurement point (N is a natural number equal to or more than one) about each of the K pieces of measurement points; an altitude storage unit that acquires altitude information on each of the K pieces of measurement points, and stores the acquired altitude information in a storage unit; an inclination calculation unit that calculates a second inclination with respect to a horizontal surface from the N-1-th measurement point to the N-th measurement point about each of the K pieces of measurement points, using K pieces of altitude information; and an extraction unit that compares the first inclination with the second inclination about each of the K pieces of measurement points, and extracts the altitude information on the measurement point high in a consistency degree of the second inclination with respect to the first inclination.SELECTED DRAWING: Figure 1

Description

  The technique disclosed in this specification relates to a technique for creating map data for an environmental map of a target area.

  For example, when the moving body autonomously moves within the target area, an environment map that stores obstacles (for example, buildings) in the target area is required. For this reason, a technique for creating an environmental map in a target area has been developed. For example, a method for creating an environmental map from environmental information acquired by a sensor is known. In this method, a moving body equipped with a sensor such as a laser range finder (LRF) is moved within a target region, and environmental information is acquired by the sensor at a plurality of measurement points on the movement path. Then, using the map data generation device, an environment map is created based on the environment information acquired at a plurality of measurement points. In addition, in such a method, a technique for improving the accuracy of the environmental map has been developed. For example, Patent Document 1 discloses a technique for improving the accuracy of an environmental map using coordinates measured by GPS. In the technique disclosed in Patent Document 1, the position of a moving body (that is, a sensor) that is moving is measured by GPS as absolute coordinates. The position of the sensor that measured the environmental information is corrected using absolute coordinates. By correcting the sensor position using the absolute coordinates, an error generated when the environmental information is measured by the sensor is reduced.

  Here, when creating a three-dimensional environmental map, elevation information is required. For example, in Patent Document 1, absolute coordinates measured by GPS are three-dimensional coordinates. In addition, as the altitude information in absolute coordinates, altitude information measured by aviation laser surveying can be used in place of the altitude information measured by GPS.

International Publication No. 2014/077684

  In the technique of Patent Document 1, in order to create a highly accurate three-dimensional environment map, the position of the sensor that measured the environment information is corrected using the three-dimensional coordinates measured by the GPS. However, when positioning with GPS, since the position of the sensor is measured from the sky, there is a problem that the latitude and longitude coordinates, which are horizontal information, are high in accuracy, but the altitude information is low in accuracy. In the aerial laser surveying, if there are buildings or trees in the vicinity, they may affect the survey results in the surrounding height direction. For this reason, the aerial laser surveying also has a problem that it cannot be said that all the altitude information in the target area is reliable.

  This specification discloses a technique for creating map data for a highly accurate three-dimensional map.

  The map data generation device disclosed in this specification generates map data for a three-dimensional map of a target area. The map data generation apparatus is K environment information for each of K measurement points (K is a natural number of 1 or more) in the target region, and indicates a relative distance to an object existing around the measurement point. An environmental information storage unit that stores environmental information, an inclination information storage unit that stores a first inclination with respect to the horizontal plane from the (N−1) th to the Nth measurement point, for each of the K measurement points, and K pieces For each of the measurement points, the elevation information storage unit that acquires the elevation information and stores it in the storage unit, and for each of the K measurement points, the first to the horizontal plane from the (N−1) th to the Nth measurement point The inclination calculation unit that calculates two inclinations using K elevation information and the first inclination and the second inclination for each of the K measurement points are compared, and the degree of coincidence of the second inclination with respect to the first inclination is determined. Extract the elevation information for high measurement points It includes an extraction unit, a.

  In the map data generating apparatus, the extraction unit extracts the elevation information about the measurement point having a high degree of matching of the second inclination with the first inclination. For this reason, even if the altitude information for each measurement point includes high accuracy and low accuracy, the altitude of the measurement point with high accuracy is obtained by comparing the second inclination with the high accuracy first inclination. Only information can be extracted. For this reason, the precision of three-dimensional map data can be made high.

The figure which shows the system configuration | structure of the map data generation apparatus which concerns on Example 1,2. The flowchart which shows an example of map data generation processing. It is a figure which shows typically the movement path | route which a moving body moves, measuring environmental information, (a) is a perspective view, (b) is a top view, (c) is a side view. 3 is a flowchart showing details of processing in step S16 in FIG. 2 according to the first embodiment. It is a figure which shows the relationship between the pitch angle measured at each measurement point, and time, (a) shows the pitch angle measured at each measurement point in time series, (b) respond | corresponds to (a). The actual elevation to be shown. It is a figure which shows the relationship between the pitch angle measured at each measurement point, time, and the relationship between altitude information and time, (a) shows the pitch angle measured at each measurement point in time series, b) shows the altitude information in a time series corresponding to (a). It is a figure which shows the relationship between the pitch angle measured at each measurement point, time, and the relationship between altitude information and time, (a) shows the pitch angle measured at each measurement point in time series, FIG. 6B shows a state where the altitude information shown in FIG. 6B is further divided every time Δt. The figure which shows an absolute coordinate typically. It is a figure which shows graph structure data typically, (a) is a perspective view, (b) is a top view, (c) is a side view. The enlarged view of the part X of FIG. 9 is a flowchart showing details of the process in step S16 in FIG. 2 according to the second embodiment. It is a figure which shows the relationship between the z coordinate of the relative coordinate of each measurement point, and a movement distance, and the relationship between the altitude information and the corresponding movement distance, (a) is the z coordinate and the movement distance of the relative coordinate of each measurement point. (B) shows elevation information in association with (a).

  The main features of the embodiments described below are listed. The technical elements described below are independent technical elements and exhibit technical usefulness singly or in various combinations, and are limited to the combinations described in the claims at the time of filing. It is not a thing.

(Characteristic 1) In the map data generation device disclosed in the present specification, the extraction unit has a group of M measurement points (M is a natural number of 2 or more and K or less) that have a positive or negative first slope and are continuous. May be extracted. According to such a configuration, a group of M measurement points that are continuous with a positive or negative first slope is extracted from the K measurement points. That is, the K measurement points are divided into a group having a positive first inclination or a group having a negative first inclination. Thereby, it becomes easy to compare the first inclination and the second inclination, and it is possible to easily determine whether or not the degree of coincidence of the second inclination with respect to the first inclination is high.

(Characteristic 2) In the map data generation device disclosed in the present specification, the extraction unit determines whether the second slope has the same sign as the first slope in the whole group of M measurement points. The degree of coincidence of two inclinations with respect to the first inclination may be determined. According to such a configuration, it is possible to easily determine the degree of coincidence of the second inclination with respect to the first inclination.

(Characteristic 3) In the map data generation device disclosed in the present specification, the extraction unit calculates a difference between the first inclination and the second inclination in the entire group of M measurement points, and the difference is equal to or less than a threshold value. Sometimes, it may be determined that the degree of coincidence of the second inclination with respect to the first inclination is high. According to such a configuration, the degree of coincidence of the second inclination with respect to the first inclination can be easily determined by calculating the difference between the first inclination and the second inclination.

(Feature 4) A map data generation device disclosed in the present specification includes a coordinate generation unit that generates three-dimensional coordinates of measurement points corresponding to elevation information extracted by an extraction unit, and K pieces of environmental information measured A measurement point node that defines the position of each measurement point, a measurement point arc that defines the relative positional relationship between adjacent measurement point nodes, a coordinate node that defines the position of a three-dimensional coordinate, and the position of the measurement point A graph structure data generation unit that generates graph structure data including a coordinate arc that defines a deviation from the three-dimensional coordinates corresponding to the position, and the graph structure data is calculated from each measurement point arc and each coordinate arc And a graph structure data optimizing unit that optimizes the graph structure data so that the sum of the error functions to be minimized. According to such a configuration, the coordinate generation unit generates only the three-dimensional coordinates corresponding to the elevation method determined to have a high degree of coincidence of the second inclination with respect to the first inclination by the extraction unit. Only coordinates with high degrees are generated. Therefore, by generating and optimizing the graph structure data from the position of the measurement point and the three-dimensional coordinates corresponding to the position, only coordinate information with high accuracy reliability is added to the graph structure data. For this reason, the accuracy of the entire three-dimensional map data can be increased.

(Feature 5) A map data generation device disclosed in the present specification uses map data in optimized graph structure data and map data using environment information measured from a position defined by the measurement point node. You may further provide the map data generation part to produce | generate. According to such a configuration, only the three-dimensional coordinates with high accuracy reliability are taken in, and based on the environmental information measured from the sensor position specified by the measurement point node in the optimized graph structure data, the accuracy is improved. High map data can be generated.

Example 1
Hereinafter, the map data generation device 10 according to the first embodiment will be described. The map data generation device 10 generates map data for a three-dimensional map of the target area. The map data generation device 10 can be configured by a computer including a CPU, a ROM, a RAM, and the like, for example. When the computer executes the program, the map data generation device 10 includes an elevation determination unit 16, an absolute coordinate generation unit 18, a graph structure data generation unit 20, a graph structure data optimization unit 22, and a map data generation unit illustrated in FIG. It functions as 24 etc. The processing of each unit 16, 18, 20, 22, 24 of the map data generation device 10 will be described in detail later.

  Further, as shown in FIG. 1, the map data generation device 10 includes an environment measurement data storage unit 12 and an altitude information storage unit 14. The environmental measurement data storage unit 12 includes environmental information in the target area measured by the environmental measurement sensor 52, attitude information of the environmental measurement sensor 52, relative position of the environmental measurement sensor 52, and two-dimensional coordinates of the environmental measurement sensor 52. Remember. As will be described later, the environment measurement sensor 52 is mounted on the moving body 50. Hereinafter, the posture information of the environmental measurement sensor 52 is also referred to as “posture information of the moving body 50”, the relative position of the environmental measurement sensor 52 is also referred to as “relative position of the movable body 50”, and the two-dimensional coordinates of the environmental measurement sensor 52 are It is also referred to as “two-dimensional coordinates of the moving object 50”. The moving body 50 measures the environment (for example, a measurement object such as a building or trees) in the target area by the environment measuring sensor 52 while moving in the target area. The environment information acquired by the environment measurement sensor 52 is stored in the environment measurement data storage unit 12 together with the posture information, the relative position, and the two-dimensional coordinates of the moving body 50 when the environment information is acquired. The altitude information storage unit 14 stores altitude information of the point where the environmental measurement sensor 52 measures the environmental information. The environmental information, posture information, relative position and two-dimensional coordinates stored in the environmental measurement data storage unit 12, and altitude information stored in the altitude information storage unit 14 will be described in detail later.

  Processing performed by the map data generation device 10 will be described with reference to FIGS. FIG. 2 is a flowchart illustrating an example of map data generation processing performed by the map data generation device 10. As shown in FIG. 2, first, the map data generation device 10 is measured from the moving body 50 at a plurality of measurement points (in this example, K (K is a natural number of 1 or more)) in the target area. The environment information, the posture information of the moving object 50 at each measurement point, the relative position of the moving object 50 at each measurement point, and the two-dimensional coordinates of the moving object 50 at each measurement point are acquired and stored in the environment measurement data storage unit 12. Store (S12).

  Here, a procedure for acquiring environment information, posture information, relative position, and two-dimensional coordinates by the moving body 50 will be described. The moving body 50 is a wheel drive type moving body. The moving body 50 includes an operating device (not shown) and is operated by a remote control such as a joystick. In the present embodiment, the moving body 50 is steered by a remote controller, but is not limited to such a configuration. For example, the moving body 50 may be moved by driving by a handle operation by an operator, or may be moved by being manually pressed by the operator.

  The moving body 50 includes an environmental measurement sensor 52 and a storage unit 54 (see FIG. 1). The environmental measurement sensor 52 measures environmental information around the mobile object 50 on which the environmental measurement sensor 52 is mounted, posture information of the mobile object 50, and position information of the mobile object 50. In the present embodiment, the environment measurement sensor 52 mounted on the moving body 50 is a 3D-LiDRA, an inertial sensor, and a GPS sensor.

  3D-LiDRA emits laser light, and measures the time until the emitted laser light is reflected by an object and returned. From the time measured by 3D-LiDRA, the distance from 3D-LiDRA to the object is measured. Further, since the direction in which the laser light is emitted from 3D-LiDRA (that is, the incident angle of the laser light reflected from the object) is known, the orientation of the object with respect to 3D-LiDRA can be determined. 3D-LiDRA can measure environmental information such as the position, posture, and shape of buildings and trees (hereinafter referred to as “measurement 62”) viewed from the measurement point. The environment information measured by 3D-LiDRA is stored in the storage unit 54. The inertial sensor is, for example, a gyro sensor, and measures posture information (for example, pitch angle) of the moving body 50. The posture information of the moving body 50 measured by the inertial sensor is stored in the storage unit 54. The GPS sensor measures the position of the moving body 50 by GPS. In the present embodiment, the position of the environmental measurement sensor 52 measured by the GPS sensor is indicated by two-dimensional coordinates based on latitude and longitude. The two-dimensional coordinates of the moving body 50 measured by the GPS sensor are stored in the storage unit 54.

  Further, the relative position of the moving body 50 at each measurement point is calculated using the environment information measured by the environment measurement sensor 52 at each measurement point. Specifically, the environment information measured by 3D-LiDRA at one measurement point (for example, the N-1th measurement point (N is a natural number between 1 and K)) and the next measurement point (for example, N pieces) By matching the environment information measured by 3D-LiDRA at the eye measurement point), the displacement of the position and orientation between them is calculated. It can be said that the displacement calculated here is the movement amount of the moving body 50 from one measurement point to the next measurement point. As a matching method, a known method such as pattern matching or ICP (Iterative Closest Point) can be used. Thus, by calculating the movement amount of the moving body 50 between each measurement point, the posture and position (that is, the relative position) of the moving body 50 at each measurement point can be calculated. In addition, the moving body 50 may include a sensor (for example, an encoder) that detects the rotation angle of the wheel. By detecting the rotation angle of the wheel with this sensor, the moving direction and the moving amount of the moving body 50 can be calculated. Based on the calculated moving direction and moving amount of the moving body 50 and the posture information of the moving body 50 measured by the inertial sensor, the moving body 50 has the posture and position (that is, relative position) of the moving body 50 at each measurement point. Position) can be calculated.

  In the storage unit 54, the posture information, the relative position, and the two-dimensional coordinates of the moving body 50 at the measurement point are stored in time series in association with the environment information measured by the environment measurement sensor 52 at each measurement point. In the present embodiment, the environment information measured by the environment measurement sensor 52, the posture information and two-dimensional coordinates of the moving body 50, and the calculated relative position of the moving body 50 may be collectively referred to as “environment measurement data”. In this embodiment, 3D-LiDRA, inertial sensor, and GPS sensor are used as the environmental measurement sensor 52, but the configuration is not limited to this. Any configuration may be used as long as environmental information can be measured from each measurement point, and posture information, relative position, and two-dimensional coordinates of the moving body 50 at each measurement point can be acquired or calculated. For example, the environmental measurement sensor 52 may be a laser distance sensor, a single-application or compound-eye camera system, an atmospheric pressure sensor, an inclinometer, or the like.

  FIGS. 3A to 3C schematically show a moving path 60 on which the moving body 50 moves while measuring environmental information by the environment measuring sensor 52. As shown in FIGS. 3A to 3C, the moving path 60 is substantially flat from the start point P0 to the point P1, and is up from the point P1 to the point P2. And it is down from the point P2 to the end point P3. In addition, there are a measurement object 62a and a measurement object 62b that protrude in the height direction (the z direction in FIG. 3A) from the movement path 60 in the vicinity of the ascending section between the points P1 and P2. For example, the measurement object 62a is a building, and the measurement object 62b is trees. The moving body 50 measures the environment information, the posture information of the moving body 50, and the two-dimensional coordinates at the K measurement points while moving along the movement path 60, and calculates the relative position of the moving body 50 at each measurement point. calculate. The measured environment information, posture information and two-dimensional coordinates of the moving body 50, and the calculated relative position of the moving body 50 are stored in the storage unit 54 together with time-series information at the time of measurement.

  Next, the map data generation device 10 acquires elevation information for the K measurement points and stores it in the elevation information storage unit 14 (S14). In the present embodiment, elevation information is acquired for each of the K measurement points from a database maintained by the Geographical Survey Institute (hereinafter also referred to as a Geographical Survey Map) and stored in the elevation information storage unit 14.

  Here, a method for acquiring the altitude information of the measurement point from the Geographical Survey Institute map will be described. Using the Geographical Survey Map, the altitude at that point can be obtained from the latitude and longitude. The elevation of the Geographical Institute map is measured by aviation laser surveying. In aviation laser surveying, a laser scanner mounted on an aircraft emits laser light toward the ground and measures the distance from the aircraft to the ground. Moreover, the position information of the aircraft is acquired from a global satellite positioning system (GNSS) surveying instrument and an inertial measurement unit (IMU). In aviation laser surveying, ground altitude and topographic shape are measured from the distance from the aircraft to the ground and the position information of the aircraft. At this time, the measured elevation data is interpolated at intervals of 5 m in the north-south and east-west directions. In the Geographical Survey Map, when latitude and longitude are input, elevation information at that point is output. The map data generation device 10 inputs the two-dimensional coordinates stored in the environmental measurement data storage unit 12 to the Geographical Survey map, and outputs the elevation information of the position corresponding to the two-dimensional coordinates from the Geographical Survey map. The output elevation information is stored in the elevation information storage unit 14 together with the corresponding two-dimensional coordinates and time-series information when the two-dimensional coordinates are measured.

  Next, the altitude determination unit 16 extracts the altitude information determined to have high reliability from the altitude information of K measurement points stored in the altitude information storage unit 14 (S16). The elevation information acquired from the Geographical Survey Map may not match the elevation of the actual travel route 60. As described above, the altitude information of the Geographical Survey Map is measured by aviation laser surveying and interpolated. For this reason, for example, if there are objects (e.g., buildings, trees, etc.) higher than the altitude of the position around the position measured by the aviation laser surveying, those objects affect the measurement result, The altitude at the measurement point may not be measured accurately. In step S16, the altitude determination unit 16 determines whether or not the accuracy information of the K measurement points acquired from the Geographical Institute map is highly reliable, and the altitude determined to have high accuracy reliability. Extract information only.

  Here, the process of step S16 will be described with reference to FIG. As shown in FIG. 4, in the process of step S <b> 16, first, the elevation determination unit 16 determines the mobile object 50 from the posture information of the mobile object 50 stored in the environment measurement data storage unit 12 for K measurement points. The posture (that is, the pitch angle) is classified into a group that is continuously positive in time series, a group that is continuously negative, and a group that is continuously approximately 0 (S32).

  FIG. 5A is a graph showing the relationship between the pitch angle measured at K measurement points and time. In FIG. 5A, the vertical axis represents the pitch angle, and the horizontal axis represents time. Graph A shows the pitch angle of each measurement point measured by the inertial sensor in time series. FIG. 5B shows the actual altitude corresponding to FIG. 5A, and graph B corresponds to the actual altitude of the moving path 60 for each measurement point on the graph A for comparison. It is attached. When the moving body 50 moves uphill, the moving body 50 faces upward, so that the pitch angle becomes a negative value. On the other hand, when the moving body 50 moves downhill, the moving body 50 faces downward, and the pitch angle becomes a positive value. As shown in FIG. 5A, the pitch angle is substantially zero continuously from time T0 to time T1. Therefore, it can be seen that the moving body 50 has moved along the flat path from time T0 to time T1, which coincides with the flat section from the start point P0 to the point P1 shown in FIG. The altitude determination unit 16 classifies the measurement point where the moving body 50 has moved between time T0 and time T1 (that is, the measurement point between the start point P0 and the point P1 in FIG. 3) as the group G1.

  Further, the pitch angle has a negative value continuously from time T1 to time T2. Therefore, it can be seen that the moving body 50 has moved on the up route from time T1 to time T2, which coincides with the up section from point P1 to point P2 shown in FIG. The altitude determination unit 16 classifies the measurement point where the moving body 50 has moved between time T1 and time T2 (that is, the measurement point between point P1 and point P2 in FIG. 3) as group G2. Further, the pitch angle is continuously positive from time T2 to time T3, and it can be seen that the moving body 50 has moved down the route. The altitude determination unit 16 classifies the measurement point where the moving body 50 has moved between time T2 and time T3 as group G3. In this way, all the K measurement points are classified into a group in which the pitch angle is continuously positive in time series, a group in which the pitch angle is continuously negative, and a group in which the pitch angle is continuously approximately 0.

  Next, the altitude determination unit 16 classifies the altitude information stored in the altitude information storage unit 14 into the group classified in step S32 (S34). As described above, the altitude information is stored in the altitude information storage unit 14 together with the time series information when the corresponding two-dimensional coordinates are measured. The elevation determination unit 16 classifies the elevation information stored in the elevation information storage unit 14 into each group set in step S32 so as to correspond to the time series information of the measurement points classified into each group. FIG. 6A is a graph similar to FIG. 5A, and shows the pitch angle of each measurement point measured by the inertial sensor in time series. FIG. 6B is obtained by adding the altitude information stored in the altitude information storage unit 14 to the graph shown in FIG. The graph C shows the altitude information stored in the altitude information storage unit 14 in the corresponding time series order. As shown in FIG. 6B, the elevation information stored in the elevation information storage unit 14 is classified into groups G1 to G3.

  Next, the altitude determination unit 16 further divides the altitude information for each time Δt, and calculates the altitude change rate for each time Δt (S36). The altitude change rate for each time Δt is indicated by positive, negative, or substantially zero. For example, as shown in FIG. 7B, when the elevation information is divided every time Δt, the elevation information is divided into three small groups g11, g12, and g13 in the group G1. In the group G2, the elevation information is divided into five small groups g21 to g25, and in the group G3, the elevation information is divided into five small groups g31 to g35. Then, the altitude determination unit 16 calculates an altitude change rate for each small group divided every time Δt. For example, when the time for defining a small group is from time t to time t + 1, the altitude information corresponding to time t is ht, and the altitude information corresponding to time t + 1 is ht + 1, the altitude information of the small group The rate of change is calculated by (ht + 1−ht) / Δt. When the altitude change rate of the small group is calculated in this way, in the group G1, the altitude change rates of the small groups g11 to g13 are approximately 0, approximately 0, and positive in time series order. In the group G2, the change rate of the altitudes of the small groups g21 to g25 is substantially zero. Furthermore, in the group G3, the altitude change rates of the small groups g31 to g35 are approximately 0, negative, approximately 0, negative, and negative in time series order.

  Next, the elevation determination unit 16 determines the undulation state of the elevation information for each group classified in step S34 (S38). The undulation state of the elevation information is determined based on the altitude change rate of the small group calculated in step S36. Specifically, among the change rates of the altitudes of the small groups constituting each group, the largest one is determined as the undulation state of the entire group. For example, in the group G1, among the change rates of the altitudes of the three small groups g11 to g13, two are approximately 0, and one is positive, so approximately 0 is the most. For this reason, the elevation determination unit 16 determines that the undulation state of the elevation information in the entire group G1 is flat. Further, in the group G2, since the change rates of the altitudes of the five small groups g21 to g25 are all about 0, the altitude determination unit 16 determines that the undulation state of the altitude information in the entire group G2 is flat. Further, in the group G3, among the change rates of the altitudes of the five small groups g31 to g35, two are substantially 0 and three are negative, so the negative is most common. For this reason, the altitude determination unit 16 determines that the undulation state of the altitude information in the entire group G3 is down.

  Next, the elevation determination unit 16 compares the undulation state of the elevation information determined in step S38 with the undulation state determined from the pitch angle of the corresponding group, and determines whether or not the undulation state of both coincides. (S40). As described above, in step S32, the pitch angle is classified into a group that is continuously positive in time series, a group that is continuously negative, and a group that is continuously substantially zero. That is, in the group where the pitch angle is positive, the undulation state determined from the pitch angle is down, in the group where the pitch angle is negative, the undulation state determined from the pitch angle is up, and in the group where the pitch angle is approximately 0 The undulation state determined from the pitch angle becomes flat. The elevation determination unit 16 compares the undulation state determined from the pitch angle with the undulation state of the elevation information in the corresponding group for each group classified in step S <b> 32, and determines whether or not the undulation states of the two match. judge.

  For example, the group G1 is classified as a group having a pitch angle of approximately 0 in step S32. That is, in the group G1, the undulation state determined from the pitch angle is flat. In the group G1, the undulation state of the elevation information is determined to be flat in step S38. Therefore, in the group G1, the undulation state determined from the pitch angle and the undulation state of the elevation information are both flat. For this reason, the elevation determination unit 16 determines that the undulation state determined from the pitch angle matches the undulation state of the elevation information in the group G1.

  Since the group G2 is classified as a group having a negative pitch angle in step S32, the undulation state determined from the pitch angle is up. On the other hand, in the group G2, the undulation state of the elevation information is determined to be flat in step S38. For this reason, in the group G2, the elevation determination unit 16 determines that the undulation state determined from the pitch angle does not match the undulation state of the elevation information. Further, since the group G3 is classified as a group having a positive pitch angle in step S32, the undulation state determined from the pitch angle is down. In step S38, the undulation state of the elevation information is also determined to be down for the group G2. For this reason, the elevation determination unit 16 determines that the undulation state determined from the pitch angle matches the undulation state of the elevation information in the group G3.

  When it is determined that the undulation state determined from the pitch angle matches the undulation state of the corresponding group of elevation information (YES in step S40), the elevation determination unit 16 determines the elevation information classified into that group. Is extracted (S42). 7A and 7B, in the group G1 and the group G3, it is determined that the undulation state determined from the pitch angle matches the undulation state of the altitude information. Therefore, the elevation determination unit 16 extracts the elevation information classified into the group G1 and the group G3. Thereafter, the process proceeds to step S44.

  On the other hand, if it is determined that the undulation state determined from the pitch angle does not match the undulation state of the corresponding group of elevation information (NO in step S40), the elevation determination unit 16 skips step S42. Then, the process proceeds to step S44. That is, if the undulation state determined from the pitch angle does not match the undulation state of the elevation information, the elevation information classified into that group is not extracted. 7A and 7B, in the group G2, it is determined that the undulation state determined from the pitch angle does not match the undulation state of the altitude information. Therefore, the elevation determination unit 16 does not extract the elevation information classified into the group G2. The inertial sensor can accurately measure whether the pitch angle at the measurement point is positive, negative, or substantially zero. Therefore, by comparing the undulation state of the altitude information with the undulation state determined from the pitch angle, it is possible to extract only the altitude information with high accuracy reliability.

  Next, the altitude determination unit 16 determines in step S40 for all the groups (that is, determines whether the undulation state determined from the pitch angle matches the undulation state of the altitude information of the corresponding group). Is determined (S44). If the determination in step S40 is not executed for all groups (NO in step S44), the process returns to step S40, and the processes in steps S40 to S44 are repeated. On the other hand, if the determination in step S40 is executed for all groups (YES in step S44), the process in step S16 is terminated and the process proceeds to step S18 in FIG.

  When the process of step S16 is completed, the absolute coordinate generation unit 18 uses the elevation information extracted in step S16 and the corresponding two-dimensional coordinates to obtain three-dimensional coordinates (hereinafter, these three-dimensional coordinates are also referred to as absolute coordinates). Generate (S18). Therefore, in step S18, the absolute coordinate generation unit 18 generates absolute coordinates only for the elevation information extracted in step S16, and does not generate absolute coordinates for the elevation information not extracted in step S16. That is, in this embodiment, absolute coordinates are generated from the elevation information classified into the group G1 and the group G3 in step S34, and no absolute coordinates are generated for the elevation information classified into the group G2 in step S34. The two-dimensional coordinates correspond to latitude and longitude, and the latitude and longitude can be converted into units of meters by converting into geocentric orthogonal coordinates with the center of the earth ellipsoid as the origin.

  FIG. 8 schematically shows the absolute coordinates generated in step S18. As shown in FIG. 8, in the region R1, absolute coordinates generated from the altitude information classified into the group G1 and the corresponding latitude and longitude are shown. Further, in the region R3, absolute coordinates generated from the altitude information classified into the group G3 and the corresponding latitude and longitude are shown. On the other hand, absolute coordinates are not shown in the region R2. This is because the altitude information classified into the group G2 was not extracted in step S18. Therefore, when the absolute coordinates generated in step S18 are arranged in chronological order, an area that does not exist continuously (area R2 in FIG. 8) exists.

  Next, the graph structure data generation unit 20 generates graph structure data from the posture information and relative position of the moving body 50 stored in the environment measurement data storage unit 12 and the absolute coordinates generated in step S18 (S20). ). As shown in FIGS. 9 and 10, the graph structure data includes a measurement point node 32, a measurement point arc 34, a coordinate node 36, and a coordinate arc 38. The measurement point node 32 defines the posture and position (relative position) of the moving body 50 at the K measurement points stored in the environment measurement data storage unit 12. The measurement point arc 34 defines the relative positional relationship of the moving body 50 between the measurement point nodes 32 adjacent in time series. The measurement point arc 34 represents a restriction on each posture of the two measurement point nodes 32 at both ends of the measurement point arc 34. The measurement point arc 34 can be calculated by matching environmental information measured by the environmental measurement sensor 52 (specifically, 3D-LiDRA) at each measurement point. The coordinate node 36 defines the position of the absolute coordinate generated in step S18. The coordinate arc 38 defines a shift between the position of the measurement point node 32 and the position of the coordinate node 36 corresponding to the position.

  Next, the graph structure data optimizing unit 22 optimizes the graph structure data generated in step S20 (S22). Specifically, the graph structure data optimization unit 22 optimizes the graph structure data so that the sum of error functions calculated based on the measurement point arc 34 and the coordinate arc 38 in the graph structure data is minimized. To do. The error function can be generated using a known method used in the Graph-based SLAM technology, and the optimization method is also known such as the least square method such as the steepest descent method or the conjugate gradient method. This method can be used. The graph structure data optimization unit 22 calculates an error function from all the arcs (measurement point arc 34 and coordinate arc 38) in the graph structure data, adds the calculated error functions, and then calculates the error function. Optimize so that the sum is minimized. The relative position of the moving body 50 calculated from the environment information measured by the environment measuring sensor 52 is highly accurate over a short distance, but an error is accumulated in the entire moving path 60. By optimizing the graph structure data to which the coordinate arc 38 is added, the graph structure data incorporating the absolute coordinates can be optimized. Thereby, the accuracy of the relative position of the moving body 50 can be increased in the entire moving path 60.

  Next, the map data generation unit 24 generates environment map data of the target area based on the environment information stored in the environment measurement data storage unit 12 and the graph structure data optimized in step S22 (S24). . Specifically, the map data generation unit 24 applies to the measurement point node 32 (hereinafter also referred to as the optimized measurement point node 32) constituting the graph structure data optimized by the graph structure data optimization unit 22. Based on the environmental information measured from the position corresponding to the measurement point node 32, environmental map data is generated. The map data generation unit 24 applies the position and orientation of the moving body 50 defined by the optimized measurement point node 32 to the environment information for each measurement point node 32. Then, the map data generation unit 24 matches a plurality of environment information to which the position and orientation of the moving body 50 defined by the optimized measurement point node 32 are applied using a known method such as ICP and the environment. Generate map data. Thus, by generating the environment map data based on the graph structure data optimized by the graph structure data optimization unit 22, the accuracy of the entire environment map data can be increased.

  In the present embodiment, the altitude information storage unit 14 stores altitude information of all K measurement points, but is not limited to such a configuration. What is necessary is just to be able to compare the undulation state of the elevation information and the undulation state determined from the pitch angle. For example, the elevation information storage unit 14 may store only the elevation information of the point selected from the K measurement points. .

  In this embodiment, the graph structure data optimizing unit 22 optimizes the graph structure data so that the sum of error functions calculated based on the measurement point arc 34 and the coordinate arc 38 is minimized. It is not limited to such a configuration. For example, an arc (hereinafter also referred to as a third arc) that defines the relative position and orientation of the moving body 50 between two measurement point nodes 32 that are not adjacent in time series is added to the graph structure data generated in step S20. May be. The two measurement point nodes 32 selected to define the third arc have a relatively short distance between the moving bodies 50 defined by the two measurement point nodes 32, and the two measurement point nodes 32 are. The time when the environmental information is measured from the moving body 50 defined by Then, the graph structure data optimization unit 22 may optimize the graph structure data so that the sum of the error functions calculated based on the measurement point arc 34, the coordinate arc 38, and the third arc is minimized. . By adding the third arc to the graph structure data and optimizing the graph structure data, the accuracy of the entire environmental map data can be further increased.

(Example 2)
In the first embodiment, when the highly reliable altitude information is extracted in step S16, the pitch angle of the moving object 50 stored in the environment measurement data storage unit 12 is used. It is not limited. It is only necessary to compare the inclination calculated from the altitude information acquired from the Geographical Survey map with the inclination of the measurement point. For example, the inclination of the measurement point may be calculated from the three-dimensional position fluctuation amount of the moving object 50. In the present embodiment, the posture information of the moving body 50 stored in the environment measurement data storage unit 12 and the altitude information extraction process in step S16 are different from those in the first embodiment, and other configurations are omitted. It is the same. Therefore, the description of the same configuration as that of the map data generation device 10 of the first embodiment is omitted.

  In the present embodiment, the environment measurement data storage unit 12 stores posture information of the moving body 50 measured by three-dimensional gyro odometry. The moving body 50 is equipped with a gyro sensor, and measures the amount of rotation about the X axis, the Y axis, and the Z axis by the gyro sensor. The moving body 50 calculates the posture fluctuation amount of the moving body 50 from the rotation amounts around the X axis, the Y axis, and the Z axis. When the moving body 50 moves, a new posture is updated based on fluctuations in the posture before the movement (for example, the posture at time t-1) and the posture after the movement (for example, the posture at time t). Further, the moving body 50 calculates the translation amount from the rotation amount of the wheels of the moving body 50. The moving body 50 calculates the three-dimensional position fluctuation amount of the moving body 50 from the updated posture and the calculated translation amount, and the three-dimensional position fluctuation amount of the moving body 50 is calculated using relative coordinates (x, y). , Z). The storage unit 54 stores the relative coordinates of the moving body 50 at the measurement points in time series in association with the environment information measured by the environment measurement sensor 52 at each measurement point.

  Next, the process of step S16 in a present Example is demonstrated with reference to FIG. As shown in FIG. 11, first, the altitude determination unit 16 has positive slopes in time series that are calculated from relative coordinates stored in the environmental measurement data storage unit 12 for K measurement points. The group is classified into a group, a group that is continuously negative, and a group that is continuously approximately 0 (S52).

  FIG. 12A is a graph showing the relationship between the z coordinate of the relative coordinates and the moving distance of the moving body 50. In FIG. 12A, the vertical axis indicates the z coordinate of the relative coordinates, and the horizontal axis indicates the movement distance. The graph D shows the z coordinate of the relative coordinates and the moving distance of the moving body 50. As shown in FIG. 12A, when the relationship between the z coordinate of the relative coordinates and the movement distance is shown, the inclination of the moving body 50 at each measurement point can be calculated. The altitude determination unit 16 detects an inflection point of the inclination of the moving body 50 and classifies the inflection point into a group in which the inclination is continuously positive, a group in which the inclination is negative, and a group in which the inclination is substantially zero. In FIG. 12A, the slope of the graph D is substantially 0 when the moving distance is from 0 to d1. Therefore, a measurement point located between the movement distance of 0 and d1 is classified as a group G1. In addition, the slope of the graph D is positive when the moving distance is from d1 to d2. Therefore, measurement points located between the movement distances d1 and d2 are classified as group G2. Further, the slope of the graph D is negative when the moving distance is from d2 to d3. Therefore, measurement points located between the movement distances d2 and d3 are classified as group G3.

  Next, the altitude determination unit 16 classifies the altitude information stored in the altitude information storage unit 14 into the group classified in step S32 (S54). Note that the process of step S54 is substantially the same as the process of step S34 of the first embodiment, and thus detailed description thereof is omitted. FIG. 12B is a graph showing the altitude information corresponding to FIG. For comparison, the graph E shows the actual elevation of the travel route 60 in association with the graph D in FIG. The graph F shows the elevation information stored in the elevation information storage unit 14 in association with the graph D in FIG. As shown in FIG. 12B, the elevation information stored in the elevation information storage unit 14 is classified into groups G1 to G3.

Next, the elevation determination unit 16 calculates the average gradient of the z coordinate of the relative coordinate for each group classified in step S52 (S56). As shown in FIG. 12A, for example, in the group G2, the z coordinate of the movement distance d1 is z _i , the z coordinate of the movement distance d2 is z _{i + 1} , and the distance between the movement distances d1 and d2 is , ΔD _j . In this case, the average gradient of the z coordinate of the group G2 is calculated as (z _{i + 1} −z _i ) / ΔD _j . Similarly, in the group G3, when the z coordinate of the movement distance d2 is z _{i + 1} , the z coordinate of the movement distance d3 is z _{i + 2} , and the distance between the movement distances d2 and d3 is ΔD _{j + 1} , the group G3 The average gradient of the z coordinate is calculated as (z _{i + 2} −z _{i + 1} ) / ΔD _{j + 1} .

Next, the elevation determination unit 16 calculates an average gradient of the elevation information for each group classified in step S54 (S58). As shown in FIG. 12 (b), for example, in the group G2, the altitude of the moving distance d1 is _{h i,} elevation movement distance d2 is assumed to be _{h i + 1.} In this case, since the distance between the movement distances d1 and d2 is ΔD _j as in step S56, the average gradient of the elevation information of the group G2 is calculated as (h _{i + 1} −h _i ) / ΔD _j. . Similarly, in the group G3, when the altitude of the moving distance d2 is h _{i + 1} and the altitude of the moving distance d2 is h _{i + 2} , the distance between the moving distances d2 and d3 is ΔD _{j + 1.} The average gradient of the information is calculated as (hi _{+ 2−} hi _{+ 1} ) / ΔDj _{+ 1} .

  Next, the elevation determination unit 16 calculates the difference between the average gradient of the z coordinate calculated in step S56 and the average gradient of the elevation information calculated in step S58 for each group classified in step S52 and step S54. It is determined whether or not the absolute value of the difference is within a threshold value (S60). In this embodiment, the threshold value is 1%, but the present invention is not limited to such a configuration. The threshold value is not particularly limited and can be appropriately changed and used. When the absolute value of the difference between the average gradient of the z coordinate and the average gradient of the altitude information is within the threshold value (in the case of YES in step S60), it is determined that the gradients of both are substantially the same. That is, it is determined that the altitude information classified into this group has high accuracy reliability. Therefore, in this case, the elevation determination unit 16 extracts the elevation information classified into the group (S62). On the other hand, when the absolute value of the difference between the average gradient of the z coordinate and the average gradient of the elevation information is not within the threshold value (NO in step S60), it is determined that the gradients of the two do not match. That is, it is determined that the altitude information classified into this group has low accuracy reliability. Therefore, in this case, step S62 is skipped.

For example, in the group G2, it is assumed that the graph E indicating the actual altitude indicates an uphill and the average gradient is about 3%. In this case, the average gradient (z _{i + 1} −z _i ) / ΔD _j of the z coordinate shown in FIG. 12A is 3%. On the other hand, as shown in FIG. 12 (b), the average slope of the altitude information _{(h i + 1 -h i)} / ΔDj has become substantially zero. Therefore, the absolute value of the difference between the average gradient of the z coordinate and the average gradient of the altitude information is 3%. That is, in the group G2, since the absolute value of the difference between the average gradient of the z coordinate and the average gradient of the elevation information is not within 1%, the elevation information classified in this group is determined to have low accuracy reliability. The Therefore, step S62 is skipped and the altitude information classified into the group G2 is not extracted.

Further, in the group G3, it is assumed that the graph E indicating the actual altitude indicates the downward direction, and the average gradient is about −2.5%. In this case, the average gradient (z _{i + 2} −z _{i + 1} ) / ΔD _{j + 1} of the z coordinate shown in FIG. 12A is −2.5%. Then, as shown in FIG. 12B, the average slope (hi _{+ 2−} hi _{+ 1} ) / ΔD _{j + 1 of} the elevation information is also decreased as a whole. For example, the average slope of the elevation information is −3%. Suppose. Then, the absolute value of the difference between the average gradient of the z coordinate and the average gradient of the altitude information is 0.5%. That is, in the group G3, since the absolute value of the difference between the average gradient of the z coordinate and the average gradient of the elevation information is within 1%, the elevation information classified into this group is determined to have high accuracy reliability. The Accordingly, the process proceeds to step S62, and the altitude information classified into the group G3 is extracted.

  Next, the altitude determination unit 16 performs the determination in step S60 for all groups (that is, determination of whether or not the absolute value of the difference between the average gradient of the z coordinate and the average gradient of the altitude information is within a threshold). It is determined whether or not (S64). If the determination in step S60 is not executed for all groups (NO in step S64), the process returns to step S60, and the processes in steps S60 to S64 are repeated. On the other hand, when the determination in step S60 is executed for all groups (YES in step S64), the process proceeds to step S18 in FIG. Thus, even if the method of the present embodiment is used, altitude information with high accuracy reliability can be extracted from the altitude information stored in the altitude information storage unit 14. For this reason, the precision of three-dimensional map data can be made high.

  Points to be noted regarding the map data generation device 10 described in the embodiment will be described. The environmental measurement data storage unit 12 of the embodiment is an example of an “environment information storage unit” and an “inclination information storage unit”, and the elevation determination unit 16 is an example of an “inclination calculation unit” and an “extraction unit”. The absolute coordinate generation unit 18 is an example of a “coordinate generation unit”.

  As mentioned above, although the specific example of the technique disclosed by this specification was demonstrated in detail, these are only illustrations and do not limit a claim. The technology described in the claims includes various modifications and changes of the specific examples illustrated above. The technical elements described in this specification or the drawings exhibit technical usefulness alone or in various combinations, and are not limited to the combinations described in the claims at the time of filing.

10: Map data generation device 12: Environmental measurement data storage unit 14: Elevation information storage unit 16: Elevation determination unit 18: Absolute coordinate generation unit 20: Graph structure data generation unit 22: Graph structure data optimization unit 24: Map data generation Unit 50: Moving object 52: Environmental measurement sensor 54: Storage unit

Claims (6)

An apparatus for generating map data for a three-dimensional map of a target area,
K pieces of environment information for each of K measurement points (K is a natural number of 1 or more) in the target area, the environment information indicating a relative distance to an object existing around the measurement points. An environmental information storage unit for storing;
An inclination information storage unit for storing a first inclination with respect to a horizontal plane from the (N−1) th to the Nth measurement point for each of the K measurement points;
For each of the K measurement points, an elevation information storage unit that acquires elevation information and stores it in the storage unit;
For each of the K measurement points, an inclination calculation unit that calculates a second inclination with respect to the horizontal plane from the (N-1) th to the Nth measurement point using the K altitude information;
An extraction unit that compares the first inclination and the second inclination for each of the K measurement points and extracts the elevation information for the measurement points having a high degree of matching of the second inclination with the first inclination; A map data generation device comprising:   The map data generation device according to claim 1, wherein the extraction unit extracts a group of M measurement points (M is a natural number greater than or equal to 2 and less than or equal to K) having the first slope that is positive or negative. .   The extraction unit determines the degree of coincidence of the second inclination with respect to the first inclination depending on whether the positive / negative of the second inclination coincides with the positive / negative of the first inclination in the entire group of M measurement points. The map data generation device according to claim 2 which judges.   The extraction unit calculates a difference between the first inclination and the second inclination over the whole group of M measurement points, and the first inclination of the second inclination when the difference is equal to or less than a threshold value. The map data generation device according to claim 2, wherein it is determined that the degree of coincidence with is high. A coordinate generation unit that generates three-dimensional coordinates of a measurement point corresponding to the elevation information extracted by the extraction unit;
A measurement point node that defines the position of each of the K measurement points at which the environmental information is measured, a measurement point arc that defines the relative positional relationship between adjacent measurement point nodes, and the position of the three-dimensional coordinate A graph structure data generation unit for generating graph structure data, comprising: a coordinate node that defines a coordinate arc that defines a shift between a position of the measurement point and a three-dimensional coordinate corresponding to the position;
The graph structure data further includes a graph structure data optimization unit that optimizes the graph structure data so that a sum of error functions calculated from each measurement point arc and each coordinate arc is minimized. The map data generation apparatus as described in any one of 1-4.   The map data generation part which generates map data using the environmental information measured from the measurement point node in the optimized graph structure data, and the position which a measurement point node prescribes is provided. Map data generator.

Images